[Bug 253968] 13.0-BETA4: UFS with journaled soft-updates is three times SLOWER than before

[bugzilla-noreply at freebsd.org]

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=253968

--- Comment #8 from Mark Millard <marklmi26-fbsd at yahoo.com> ---
(In reply to Mark Millard from comment #7)

12.2-RELEASE on the OverDrive 1000 this time,
again operating on the SSD:

# time -l make extract
===>   firefox-85.0.2,2 depends on file: /usr/local/sbin/pkg - found
===> Fetching all distfiles required by firefox-85.0.2,2 for building
===>  Extracting for firefox-85.0.2,2
=> SHA256 Checksum OK for firefox-85.0.2.source.tar.xz.
      120.92 real        55.87 user        61.93 sys
     76520  maximum resident set size
        27  average shared memory size
        10  average unshared data size
       127  average unshared stack size
     27241  page reclaims
         6  page faults
         0  swaps
     20866  block input operations
      6362  block output operations
         0  messages sent
         0  messages received
         4  signals received
     21852  voluntary context switches
      3121  involuntary context switches

# time -l rm -fr work
       12.83 real         0.37 user         8.17 sys
      3916  maximum resident set size
         7  average shared memory size
         7  average unshared data size
       127  average unshared stack size
       934  page reclaims
         0  page faults
         0  swaps
         5  block input operations
       348  block output operations
         0  messages sent
         0  messages received
         0  signals received
       581  voluntary context switches
       736  involuntary context switches

13.0-BETA4 looks a somewhat better for this on the extract.
Of course, all my examples are SSDs or Optanes, nothing
spinning. I'm unsure how that or other context variations
would compare. But now you have some examples for
comparison.

The early Feb. main (14) had much bigger differences between
user+sys vs. real than 13.0-BETA4 or 12.2-RELEASE did. But
that may well be expected given the huge "block output
operations" count differences.

It does not appear that I've managed to run into the
13.0-BETA4 problem(s) via these specific types of
experiments for UFS2+SU (no J).

-- 
You are receiving this mail because:
You are the assignee for the bug.